Making Intelligence Analysis Responsive to Policy Concerns
CIA HISTORICAL REVIEW PROGRAM
RELEASE IN FULL
22 SEPT 93
A request from the consumer
MAKING INTELLIGENCE ANALYSIS RESPONSIVE TO POLICY CONCERNS
Lt. Colonel James Murphy and Dr. K. Wayne Smith
Should the intelligence analyst stay aloof from issues that - seize the policy makers he supports? Here are two users of finished intelligence, both former members of the National Security Council staff, who say "no." They collaborated to write a speech on this theme which Col. Murphy delivered before a gathering of intelligence officers (the Intelligence Forum) in May 1971. An abridged text of the speech follows.
Let me begin with some observations about intelligence support for the preparation of National Security Study Memoranda — NSSMs for short. This process is a systematized procedure by which the President directs the attention of the bureaucracy to national security issues in which he has an interest or needs to make a decision. It is a means of mobilizing the intellect and energy of the government and focusing them on major foreign policy issues.
The personnel and informational resources needed to address these issues are abundantly available within the government, but they tend to be fragmented among the various agencies including the intelligence community. The information and expertise required for rational decision making are interagency in nature. No single department or agency has a comer on the foreign policy market, and that goes for the intelligence input as well.
Moreover, the NSSM process gives those who have the responsibility for implementing and supervising the execution of foreign policy an opportunity to participate in the formulation of those policies. And I need not remind you that good intelligence information is essential, not only for making sound policy decisions, but also as an essential ingredient for judging the performance of policy.
The NSSM process is founded upon several principles which have a direct bearing on the intelligence response. The first is creativity. In a world of onrushing and complex change, we cannot be content with familiar ideas, or assume that the future will be merely a projection of the present and the past.
In March 1971, when the President wanted an analysis of the impact of LAMSON 719 on the North Vietnamese logistic capability to support various types and levels of warfare against South Vietnam in the future, we found that the currently available intelligence data on supply flows was simply inappropriate to measure the capabilities of the enemy. The data had been structured to portray the magnitude of the enemy effort in terms of tonnage, and the results of our interdiction operations in terms of truck kills, but without relation to the effects on future levels of enemy activity.
Within the NSC Staff, a new analytical format was created by which the existing data could be structured to indicate the implications for North Vietnamese military capabilities. The CIA took this new analytical format and produced an outstanding study that was of direct use to the President in assessing the significance of LAMSON 719 and assisted him in making policy decisions concerning our future role and presence in Vietnam.
We have found in the NSC system that the initial focus on the broader questions of national objectives and purposes stimulates discovery of new perspectives, and produces innovative approaches to the specific issues involved. I urge you to resist the strong temptation to fulfill the immediate demands of the moment with familiar standard operating procedures. Rather, reflect upon the issues in their policy context to determine the logical relationships between the informational requirements and the issues, and then decide what information needs to be presented and how.
The second principle of the NSSM process is the quest for accurate, complete, and relevant factual information. Too often the process of policy making has been impaired or distorted by incomplete information and by disputes within the government which resulted from a lack of common appreciation of the relevant facts. It is an essential function of the NSSM process to bring together all of the agencies of government concerned with foreign affairs to elicit, assess, and present to the President and the Council all the pertinent knowledge available and necessary to make policy that is thoroughly grounded in, and relevant to, the facts. It is on this rock that the intelligence community sometimes founders.
Information is your stock and trade, and the inventory is jealously guarded and highly regarded by several proprietors. Sometimes, the search for facts degenerates into a squabble about who is right, rather than what is right. I see enough documents from various intelligence agencies to know that rarely does any one agency possess all the relevant facts. Indeed, it would seem to me that the structure of the intelligence community is such that no one agency could possibly have all the facts. However, even with this conceded, I still hear disputes about who has the better information, and invariably each proponent claims that his data are best. The assessment of LAMSON 719 turned at one point into just such a hassle. The President, for whom we all work, couldn't care less about whether one agency's or another agency's sources are better truck counters. The agencies involved could have served the requirements of policy making better if they had analyzed their own information for trends and sought to establish meaningful correlations and ranges of estimates encompassing both types of data.
It took some methodological head knocking, but we were finally able to focus on what was right rather than who was right, and this was done by combining intelligence resources into meaningful patterns and relationships conducive to policy making. I realize that intelligence producers suffer peptic ulcers whenever they're asked for "the facts." It is in these moments that they have to be honest with themselves about the reliability of their sources, and the accuracy and completeness of their information. Usually, the pieces of information far outnumber the assured facts.
The phrase "true facts" is neither a linguistic nor logical redundancy to intelligence officers. Certainly policy makers, however, cannot "wait until all the facts are in." That is the job for the historians among your grandchildren. The challenge to intelligence is to be able to see through the glass darkly, here and now.
The intelligence community has a natural and understandable inherent conservatism, sometimes bordering on reluctance to commit itself to declarations of fact. Indeed, a private jargon, little understood by the consumer, hedges such statements. Estimated, believed, probably, possibly, may, and might are examples of this special vocabulary. In fact, it reminds me of one of the first meetings I attended in the Pentagon where I heard a distinguished Assistant Secretary of Defense stand up and pound the table to insist that might was too strong a word to use in a particular sentence under discussion! I don't know what is weaker than might. But the policy maker must make decisions in a context of nebulous visibility. The intelligence community must do all it can, even at the occasional risk to its own credibility, to give the decision maker as much clear insight as possible into the facts, and to make clear the hedges, risks, and consequences of error. Too often, the hedging seems self-protective, rather than informative. And it is just as important for the intelligence community to be forthright in identifying what we do not know, and to assess the consequences for policy of this ignorance. Risk and uncertainty are the environment of policy making, and the intelligence community must share and seek to relieve some of this burden. The "true facts" are that impossible dream. But relevant information and a rational, objective assessment of it remain the most essential intelligence responses to the NSSM decision-making process.
A third principle of the NSSM process is the provision of a full range of feasible options. The President's leadership cannot consist merely in being confronted with a bureaucratic consensus that leaves him no option but acceptance or rejection of a single proposal, without any way of knowing what alternatives exist. The NSSM system is designed to ensure that clear policy choices reach the top so that various positions and alternatives can be debated fully. The NSSM system ensures that all agencies involved receive a fair hearing before decisions are made. Interagency participation begins with the working groups that draft the papers and continues right up through the review process, all the way to the National Security Council itself. Legitimate alternative positions reach the President without dilution. Differences are clearly identified and defended, rather than being muted or buried in bureaucratic waffle and log rolling.
These features give the President confidence that his choices are genuine, and enable him to put his own stamp on policy by the act of decision. I commend to you this principle of a full range of options in the intelligence response to the NSSM process. Much of the intelligence we all deal with is subject to varying interpretations. It is not sufficient simply to forward information. The information must be assessed and interpreted. We're all seasoned bureaucrats and well aware of the pitfalls of bargaining for consensus. But differences of view are legitimate, and the President should be aware of them and the impact on his options for choice. Certainly where agreement is possible, it should be so stated. But the ambiguity of many of the so-called facts leaves large openings for varying but reasonable interpretations.
The significance of data to the policy maker must be made clear. The NSC should be able to rely upon the intelligence community for these interpretations. Moreover, the NSC should be made aware of differences and the reasons for them. It is not enough simply to point out that disagreements exist. They must be explained, not only in terms of the substance, but also in terms of the implications for policy choices. If the intelligence community avoids this responsibility, then others will move in to fill the vacuum.
For years, the National Intelligence Estimates had epitomized this problem. They hid as much as they revealed, if not more. In 1971, with some gentle but persistent prompting by the NSC staff, some of the NIE's were truly outstanding, particularly with regard to strategic forces. One reason is that these alternative assessments and the areas of controversy were explicitly stated with accompanying rationale. This new feature has resulted in a quantum jump in the value and utility of the NIE's for the policy maker. Reasoned alternatives give the President a basis for rational choice, rather than intuitive chance, in his policy decisions.
Crisis anticipation is the fourth principle of the NSSM system. The better prepared we are in terms of foreknowledge and options, the more we can be the master rather than the slave of events once a crisis breaks. Certainly, we cannot anticipate fully the timing and course of a possible crisis. But we can take actions to help ensure that we have asked the right questions in advance, that we have developed and explored our options, and have thought through the implications of alternative responses. The intelligence community is often the first to receive the tentative signals of impending crisis. While one flower does not make a spring, its appearance should not simply be marked by a footnote. The intelligence community should develop a discriminating instinct for crisis. I should add that this instinct should not be reined in by institutional biases and values which dull the senses and force-fit data into preconceived patterns.
I remind you of the Czechoslovak crisis and how stunned many were when it turned out to be something more than just another round of summer maneuvers. This delicate instinct for crisis requires that the intelligence producer have a reporter's nose for the news. He must be policy-oriented, be free to state his opinions with their rationale, and not be constrained simply to writing copy to fill out a daily bulletin. The intelligence producer must be more than a middleman between the collector and the consumer. He must evaluate his raw material, assess the significance, and relate it to policy.
A fifth principle of the NSSM process is systematic analysis. Policy cannot be allowed to be merely the result of ad hoc piecemeal tactical decisions forced as kneejerk reactions to the immediate pressures of events. A policy must be considered in the whole context of the situation and our national interest. The interagency nature of the NSC system assures that all relevant aspects of a problem are considered in formulating policy choices.
Now, analysis is a systematic way of thinking, a manner of approaching problems in an innovative, thorough and objective way. It requires the orderly juxtaposition of facts and values in order to make reasoned judgments in decision making. Analysis is not an occult science that produces immutable truth. It is more of an intellectual art that seeks to illuminate problems and sharpen judgment. I will be the first to concede, on the basis of several years of painful experience, that even good analysis does not necessarily bring on right decisions. But I'm willing to take my chances that a decision rationally arrived at is more likely to be right than an irrational one. Let me hasten to add that analysis does not solve problems itself. It is not a substitute for imagination, leadership, or wisdom. It brings out the bad news as well as the good. It does not make decisions for you. But it does serve to discriminate between choices, separating the knowable from the unknowable, the better from the worse, the patently wrong from the approximately right. The decision maker is thereby able to focus his attention on the issues requiring the application of his experience and values and the exercise of his judgment.
I sometimes fear that the intelligence community shortchanges its own analytical efforts and shuns participation in such efforts on an interagency basis. Too often the so-called intelligence input is simply a dry catalog of information lacking focus on the policy matters at issue. While the intelligence community is not charged with making policy, it does have the responsibility to participate in policy making. This responsibility does not stop with the mere reportage of the estimated facts. The intelligence input should include an analysis of the data in relation to policy options. That the intelligence community is capable of providing these analyses oriented to policy making has been amply demonstrated in its contribution of the NSC Verification Panel's work in preparing the SALT negotiating options.
The most crucial area for intelligence analysis is in the muddy field of enemy intentions. The usual response is to sidestep the issue, with the claim that the intelligence analyst can only provide assessments of capabilities. Capability analysis is essential to policy making, but someone has to make an educated guess as to intentions. It is not enough to know what the enemy could think. What he is more likely to think is even more important to the policy maker. Analysis offers a systematic and rational method for seeing through that glass darkly — even if only with one eye in a fog.
Finally, the NSSM system provides a means by which policy implementation can be reviewed, coordinated, and supervised. Once more, the intelligence community can provide some of the eyes and ears to detect the progress of policy implementation and the pitfalls and dangers that loom up along the way. It is false to claim that these are the tasks of the so-called "operators" — that the intelligence job is done when the policy decisions are made. The intelligence function carries through the entire policy process from inception to conclusion, which includes monitoring implementation.
Once again, too, this monitorship needs to be conducted with at least one eye on the effect for policy. You must be not over-fascinated with information for its own sake. To be useful, information must be related to policy — past, present, and future.
In summary, the intelligence response to the NSSM process has been a reflection of the strengths and weaknesses of the intelligence community itself. In general, responsiveness in providing needed information has been good, though prodding has sometimes been required to elicit that information in a form more meaningful to the decision maker. I would say that the greatest improvement needed in the intelligence community is for it to begin anticipating the needs of the policy makers, and to take the initiative in providing information structured to these needs.
Let me illustrate!
The other day I sat in an intelligence briefing on the results of some recently acquired information. A fellow NSC policy analyst in the audience told the briefers that over the past six months the intelligence community had been reporting a great number of discrete bits of information regarding the quantitative increase of the Soviet presence in Egypt. He reminded the briefers that the Arab-Israeli confrontation was still a matter of national security interest, that there was at least continuing talk of possible agreements, and that U.S. interests were involved. He suggested that, in light of the disconnected information flow, it was about time we had an analytical summary that would bring us up to date on the situation and reveal the implications for our policy. The intelligence officer's supervisor was there and he agreed that this was "a good idea" and that he would "see about doing it." Ladies and gentlemen, that "good idea" should have come from the intelligence community itself!
You read the newspapers. You know the issues. But do you see your opportunity? Expand your vision. Relate intelligence information and requirements to the needs of the policy makers. Don't ask me what is my need-to-know. Ask yourselves instead what needs to be known. Analyze the issues! Analyze your information! And analyze the meaning of the information in terms of policy implications.
In conclusion, the principles of the NSSM process — creativity, factual information, a full range of options, crisis anticipation, systematic analysis, and policy implementation — are good principles for gauging the intelligence response. I was asked to say what was "wrong" with that response. I have given you my views on how you can strengthen that response and magnify the value of your essential contribution to the decision-making process. I want to finish by saying that I already see many indications that the intelligence community is aware of these problems, and that the kinds of actions I have recommended are being taken.